Hi, I'm Ridhi — SWE/ML engineer building reliable AI systems.

I’m a Machine Learning Engineer at Ciroos and a 2026 Purdue CS grad. I build LLM-powered systems, agents, and reinforcement-learning pipelines — previously at Intel.

See projects Skills Email me

Bay Area, CA · Open to SWE/ML new-grad roles

Skills

🔠 Languages

Python Java C# C++ C JavaScript TypeScript SQL

🧠 ML & AI

PyTorch TensorFlow Keras scikit-learn XGBoost Stable-Baselines3 OpenVINO Pandas

🌐 Web & APIs

React React Native Next.js Node.js Tailwind CSS Express FastAPI Flask

☁️ Infrastructure

Docker Kubernetes Git/GitHub Linux AWS GCP PostgreSQL MongoDB Firebase

🔧 AI Tools & Frameworks

LangChain LangGraph LlamaIndex AutoGen FastMCP RAG HuggingFace Baseten TensorBoard

Experience

Machine Learning Engineer Intern — Ciroos

Oct 2025 – Present · Pleasanton, CA

LLM Evaluation Framework: Built an evaluation system benchmarking LLMs on latency, token use, and quality, automating “LLM-as-a-Judge” workflows in Python and Baseten to enable scalable model comparison and reporting.
Document Intelligence Systems: Engineered document-to-JSON parsing pipelines with Docling, Pydantic, and Claude agents, improving schema consistency and reducing hallucinations.
Real-Time Visual Analytics Platform: Built a serverless dashboard analytics tool to capture and store authenticated dashboard screenshots, integrating Vision-Language Models to automatically interpret charts and generate actionable insights.

Software Engineering Co-op — Intel

May 2025 – Oct 2025 · Folsom, CA

Diagnostic Copilot with LLMs: Built an AI-powered diagnostic copilot using Qwen-3B/8B models integrated with LangChain, FastMCP, and OpenVINO, reducing false-positive reports by 40% and enabling sub-1.2ms offline inference.
Reinforcement Learning Pipelines: Trained PPO and SAC agents for image-quality optimization, improving SSIM by 11% over baseline models while automating evaluation workflows with TensorBoard.
MCP Server & RAG Systems: Built a custom MCP server and retrieval-augmented generation pipeline for expert data, enabling real-time log parsing and agent-driven troubleshooting.
Research Communication & Impact: Authored an internal white paper, presented results in executive demos, and secured cross-division engagement in RL-driven ISP automation pipelines.

Data Engineer — CDC Protect

Aug 2023 – Feb 2025 · West Lafayette, IN

Data Pipeline Development: Designed and implemented an end-to-end data pipeline using Python, Beautiful Soup, and Selenium to scrape structured and unstructured data from children's hospital websites, supporting healthcare data analysis at scale.
Automated Data Processing: Transformed unstructured data into structured formats and automated keyword extraction workflows, increasing data retrieval efficiency by 30% and enabling faster analysis and reporting.
Image Classification Accuracy: Developed and trained Convolutional Neural Networks (CNNs) for image similarity analysis, improving classification accuracy by 15%.

Earlier — Research Internships

2023 · West Lafayette, IN & Remote

Quantum Mechanics Lab (SWE Intern): Automated Halide Perovskite DFT simulations in Python and applied SISSO-based regression and neural networks to predict material properties, improving accuracy by 20%.
Artificial Code Generation (Undergrad Researcher): Evaluated and benchmarked program-synthesis frameworks (AlphaCode, CodeRL, Parsel) against state-of-the-art models on real-world coding tasks.

My Projects

Run It Back

Mobile app to find, host, and join nearby pickup sports games.

React Native Expo AWS Amplify

GitHub ↗ Details →

⚖️

Agent Evaluation Platform

Evaluation platform for multi-turn Kubernetes diagnostic agents. Unified datasets, auto-grading, and dashboards tracking quality, latency, and structure conformance.

Python LLM-as-a-Judge Dashboards

Ciroos · proprietary

MCP Server

Modular diagnostic platform with local LLMs and RL for offline AI-powered analysis at Intel.

Python LangChain OpenVINO

GitHub ↗ Details →

AisleIQ

Dual-platform retail price intelligence. A consumer app that predicts your cheapest grocery trip, plus an ML pricing dashboard for vendors. Catapult 2026 hackathon.

Next.js Firebase PyTorch

GitHub ↗ Details →

CDC Protect

Web-scraping pipelines and CNNs analyzing similarity between cannabis-edible and children's candy packaging, supporting public-safety regulation.

Python TensorFlow Selenium

Details →

PassTheAux

Collaborative music queue. Groups vote on tracks with Spotify & Apple Music.

React Firebase Spotify API

GitHub ↗ Details →

View project archive

Contact

Best way to reach me is email. I’m open to SWE/ML new‑grad roles and research‑adjacent work.

ridhitamirasa@gmail.com · LinkedIn · GitHub